Selected Analytics &
Automation Results

Situation

Paper-based processes required repeated data entry, administrative coordination, and the physical handling of forms across teams.

Task

Digitize the workflow, reduce manual effort, and create a centralized source of operational data.

Action

Designed and implemented Power Apps within Microsoft Teams, connected Power Automate workflows, and used SharePoint for structured data capture and process visibility.

Result

• Approximately 500 hours of administrative effort saved
• Several thousand paper forms and pages eliminated
• Centralized digital data capture
• A more consistent, visible workflow

Situation

Recurring operational reports depended on manual database queries, file preparation, and distribution each reporting cycle.

Task

Create a repeatable reporting process that could run on schedule with less staff intervention.

Action

Developed SQL stored procedures, scheduled execution with Windows Task Scheduler, and automated preparation and distribution of reporting outputs.

Result

• Approximately 7 hours saved each week
• Roughly 350 hours of annual staff capacity potentially reclaimed, using a 50-week planning approximation
• More consistent report timing and delivery

Situation

A recurring browser-based process required repetitive navigation, data entry, and verification across the same sequence of screens.

Task

Reduce manual effort while preserving the steps and checks required by the business process.

Action

Built a Power Automate browser workflow to complete rule-based navigation and data-entry steps consistently.

Result

• Approximately 100 hours of annual manual effort saved
• Reduced repetitive navigation and data entry
• A consistent, repeatable process

Situation

Historical appointment patterns made it difficult to identify higher-risk upcoming appointments early enough for targeted intervention.

Task

Identify appointments at elevated risk of non-attendance and prioritize proactive outreach.

Action

Engineered features from appointment, demographic, payer, provider and service, claims, weather, and behavioral-history data. Developed Python and XGBoost models in Databricks, tracked evaluation with MLflow, and operationalized ranked delivery through SQL-based scheduled workflows.

Result

• Weekly ranked outreach lists for higher-risk upcoming appointments
• Targeted intervention focused on appointments with greater non-attendance risk
• Approximately $100K in estimated top-line therapy revenue opportunity identified
• Supported efforts to reduce gaps in care; the opportunity was modeled, not booked or generated revenue